You've started thinking about how AI systems understand your brand. You've read about entity profiles and earned media.
In earlier articles in this series, we've explored how AI systems build their understanding of businesses through entity profiles and independent signals. This article tackles the question that naturally follows: how do you know if any of it's working?
The instinct is to check your ranking position like you would in traditional SEO. Open ChatGPT, search for your category, see where you appear. Do it again tomorrow and check if you've moved up.
New research from SparkToro says this is the wrong instinct.
The Sub-1% Consistency Problem
In January 2026, SparkToro ran 2,961 prompts across ChatGPT, Claude, and Google AI using hundreds of volunteers. They asked the same questions repeatedly to see how consistent the recommendations were.
The findings were stark. There's less than a 1-in-100 chance any AI tool returns the same brand list twice for the same prompt. The same list in the same order appears less than 0.1% of the time.
AI recommendations are variable. Every response is unique.
This sounds alarming until you understand how these systems work. AI doesn't maintain a ranked list the way Google does. There's no position one staying fixed until something changes. Every time an AI system generates a response, it's assembling a fresh answer from a pool of brands it considers credible for the topic.
Of course the exact list is different every time.
The Consideration Set Stays Stable
What matters is what the SparkToro data showed underneath that variability.
The consideration set, the pool of brands AI draws from, is stable. Bose, Sony, Sennheiser and Apple appeared in 55 to 77% of nearly a thousand headphone responses. Every list was unique but the same brands kept showing up.
That's the bit that matters for marketing managers.
The question isn't what position am I. It's am I in the consideration set at all, and how consistently do I appear across many different ways people ask the same type of question.
AI systems build an understanding of which brands are credible for a given topic based on the accumulated pattern of mentions, references, reviews and independent signals across the web. When someone asks a question, the system draws from the consideration set, but the exact combination and order varies based on subtle differences in how the question is phrased, the context of the conversation, and the randomness built into how these models generate text.
This is why entity presence matters. The brands appearing in those SparkToro results aren't there because they optimised for a specific prompt. They're there because their web presence gives AI systems enough consistent, independent evidence to recommend them regardless of how the question gets asked.
Ahrefs found the same pattern from a different angle. Their research on 75,000 brands showed that branded web mentions had the strongest correlation with appearing in AI responses, stronger than traditional backlinks.
What Changes in Your Actual Work
You can't optimise your way into a specific position because positions don't exist in any stable sense. But you can build the kind of entity presence that earns you a place in the consideration set.
That's a fundamentally different activity.
In traditional SEO, most of the work happens on your own website. You're optimising pages, building links back to your domain, targeting specific keywords. When you shift to building entity presence, the focus moves outward. You're thinking about what the rest of the web says about you, not what you say about yourself.
In practical terms, that means making sure your Google Business Profile is complete and consistent. That your brand is being mentioned in industry publications and review sites. That your expertise is showing up in conversations on platforms like Reddit and relevant forums. That the information about your business is consistent everywhere it appears.
AI systems are specifically looking for independent signals. Your website can never be independent of you, so whilst it still matters for traditional search, it's not what gets you into the AI consideration set.
The other big change is moving from campaign thinking to cadence thinking. Traditional SEO often works in bursts. You do a keyword research project, optimise a batch of pages, build some links, then move on. Building entity presence requires a sustained, ongoing rhythm.
Recent research showed that half of all AI citations come from content published within the last 11 months, and the most commonly cited content was published yesterday. This isn't something you do once and walk away from.
Two Layers Working Together
The recency finding might sound like it contradicts the idea of building durable entity presence. It doesn't. It's two layers working together.
AI systems work on the same two layers as professional reputation. Your credentials, your track record, the body of work you've built over years, these form the foundation. But if someone asks who's doing interesting work in this space right now, they're going to mention people who are contributing, not someone who published something great three years ago and went quiet.
There's the training layer, which is the accumulated web presence built over months and years. That's your entity foundation and it shifts slowly. Then there's the retrieval layer, where AI systems pull in recent content through real-time web search to supplement what they already know. That's where recency matters.
The brands in that stable consideration set have both. They have deep entity foundations and they're consistently producing fresh signals.
You're building two things simultaneously. The long-term entity foundation through consistent information across the web, strong independent mentions, complete business profiles, and genuine expertise signals. And the recency layer through regular content that demonstrates you're actively contributing to your field right now.
Neither is sufficient on its own.
What to Measure When Rankings Don't Exist
If positions don't exist and lists change constantly, what should you be measuring month to month?
There are three things we'd recommend any marketing manager start doing, none of which require expensive tools.
First, run a monthly entity audit. This takes about 15 minutes. Search your brand name across ChatGPT, Perplexity and Google AI. Then search your brand plus your key service areas. Then search the unbranded questions your customers typically ask, the kind of thing like "best equipment hire company in Auckland" or "who specialises in franchise marketing in New Zealand."
You're looking for whether you appear at all, how accurately the AI describes you, and whether it associates you with the right topics. Do this monthly and you'll build a picture of your consideration set presence over time. Screenshot the results so you can compare.
Second, check your server logs for AI bot activity. GPTBot, ClaudeBot, PerplexityBot and others leave a trail when they crawl your site. If AI systems aren't crawling you, they're not indexing your content for retrieval. Most marketing managers aren't doing this at all, which means it's a quick win. Your developer or hosting provider can pull this data in minutes.
Third, track your brand mention frequency across the web. This is your proxy for consideration set strength. Are you being mentioned independently on review sites, in industry publications, in forum discussions? Are those mentions consistent with how you want to be positioned? Tools like Google Alerts are a free starting point, but the key is tracking the trend over time rather than any single snapshot.
⚠️ If someone tries to sell you an AI visibility tool, ask them two questions before spending anything. How many prompt runs are you basing your data on? And are you showing me visibility percentage across many runs, or a snapshot position from a single query?
If they can't answer those clearly, the SparkToro research tells you everything you need to know about how reliable their data is.
When AI Gets Your Brand Wrong
If AI systems already mention you but get some details wrong, that's a different problem to solve than if you don't appear at all.
Inconsistent information across the web is the main culprit. AI systems build their understanding of your brand from every mention they can find. Your website, your Google Business Profile, directory listings, review sites, social media, media coverage, all of it. If those sources tell slightly different stories about what you do, where you operate, or what you specialise in, the AI doesn't pick the most accurate version.
It either gets confused and says something muddled, or it picks up whichever signal is strongest, which might not be the one you want.
A common example is a business that has evolved over time. Maybe you started as a general marketing agency but now specialise in franchise networks. Your website says one thing but your old directory listings, LinkedIn company description, and industry association profiles still reflect the old positioning.
Inconsistent signals don't average out to something roughly correct. They reduce the AI's confidence in recommending you at all, or they produce a description blending your old and new positioning.
The fix is methodical but not complicated. Start with what the AI is actually saying about you. That entity audit gives you the baseline. Then work outward. Update your Google Business Profile, make sure your LinkedIn company page matches your current positioning, check industry directories, review your presence on any third-party sites that mention you.
The goal is consistency. Every independent source describing your business should be telling the same story.
You're not correcting the AI directly. You can't email ChatGPT and ask for a correction. You're correcting the web presence that the AI draws from. Fix the inputs and the outputs follow.
Where Specialist Support Adds Value
You can run the basic entity audit yourself. It takes 15 minutes and costs nothing. But there's a difference between checking what AI systems say about your brand and systematically strengthening your position across the consideration set.
A specialist agency that understands AEO can do several things you can't do efficiently on your own. They can run entity audits systematically across your full competitor set, not just your own brand. They can properly analyse bot crawl data to identify which AI systems are indexing which parts of your site and where the gaps are. They can build a strategic plan for strengthening your independent mention profile across review sites, industry publications, and relevant platforms.
They coordinate everything from Google Business Profile optimisation to earned media cadence to technical site structure in a way building entity presence whilst maintaining traditional search performance. Coordination produces results.
The agencies worth working with are the ones that can bridge both worlds. Traditional SEO isn't going away. Google still processes billions of searches daily and traditional rankings still drive traffic and revenue. But AI search is growing alongside it.
When you're evaluating whether to bring in specialist support, the question to ask is straightforward: can you show me how you're thinking about AI search visibility for our brand, and can you maintain our traditional search performance whilst building entity presence for AI? If they can't articulate an approach beyond keyword rankings, that tells you whether they're keeping pace with how search is evolving.
But here's what makes this less daunting than it sounds. For most businesses in New Zealand, the bar is lower than you'd think. We're a smaller market with less noise. The SparkToro research was largely looking at global categories with massive competition.
In a New Zealand context, or even in specific industry verticals, the consideration set is less crowded. If you're a specialist equipment hire company or a franchise accounting network, you're not competing against thousands of brands for AI attention. You might only need to be more visible and more consistently represented than a handful of direct competitors.
The Biggest Mistake Marketing Managers Are Making
Waiting. That's it.
The biggest mistake is treating AI search as something they'll get to eventually, once the tools are better, once the measurement is clearer, once someone proves the ROI. Meanwhile, their competitors are building entity presence right now, and that presence compounds over time.
The SparkToro data makes this more urgent, not less. Brands dominating AI consideration sets are there because of accumulated, consistent web presence, not because they found a clever shortcut last month. The training layer of these AI systems builds over 6 to 12 months. Every month you wait is a month your competitors are building a foundation you'll need to catch up to.
And the barrier to starting is so low that there's no good reason to wait.
Run an entity audit this week. It takes 15 minutes and costs nothing. Check what AI systems are saying about your business, whether you're appearing for the right topics, and where the gaps are. That gives you a baseline and a clear picture of what to work on.
The businesses that will be in the strongest position a year from now aren't the ones with the biggest budgets or the fanciest AI visibility tools. They're the ones that started building a consistent, accurate, independent web presence today.
In New Zealand especially, where the market is smaller and the consideration sets are less crowded, the opportunity to establish yourself now whilst competitors are still waiting is genuinely significant.
The fundamentals haven't changed. Build a genuine reputation, make sure the web accurately reflects your expertise, and maintain that presence consistently. AI search hasn't made that harder. It just made it more important.
💡 If you need help understanding the terminology used in this article, ADMATIC maintains an AEO/GEO Glossary that explains the key concepts.
Source: Rod Russell, Managing Partner, ADMATIC, 10th March 2026